from transformers import pipeline
from PIL import Image
import torch
import numpy as np
import gradio as gr

depth_estimator = pipeline(task="depth-estimation", model="Intel/dpt-hybrid-midas")

def generate_depth_image(input_image):
    out = depth_estimator(input_image)

    # resize the prediction
    prediction = torch.nn.functional.interpolate(
        out["predicted_depth"].unsqueeze(1),
        size=input_image.size[::-1],
        mode="bicubic",
        align_corners=False,
    )

    # normalize the prediction
    output = prediction.squeeze().numpy()
    formatted = (output * 255 / np.max(output)).astype("uint8")
    depth = Image.fromarray(formatted)
    return depth

demo = gr.Interface(title = "Depth Estimation of the Detected Objects in the Image - Test & Demo App by Srinivas.v..",
                    description='Upload an image that has some vivid foreground objects and submit',
                    fn = generate_depth_image, inputs=gr.Image(type='pil'), outputs=gr.Image(type='pil'))
demo.launch(share=True,debug=True)